[Lapha]

Ideally it's lifecycle carbon footprint averaged over expected power production during that time, but the numbers aren't always directly comparable and the error bars are huge to boot.

With solar, most of the carbon footprint comes from the massive energy requirements needed to produce the panels, and countries that use coal to produce panels like China have a significantly higher carbon footprint than say the EU. Then there's the matter of where you install them, panels in most of Europe will have a CO2/kWh figure similar to to panels installed in north-west Canada, while panels installed in the US will have a similar figure to panels installed in southern Europe and northern Africa. Newer panels generally have a longer lifespan and are more efficient so will have a lower number even if the total footprint stays the same. Should the albedo effect be considered with solar? It matters if you plan to cover a light desert with dark panels. How often it rains can also ironically have a difference, as rain helps clean the panels keeping them running efficiently.

With nuclear, there's the mostly fixed costs of constructing and decommissioning the plants, the ongoing cost of running and maintaining the plants, disposing of spent fuel, and with most of the cost coming from mining the fuel. To me nuclear seems a little more straight forward to calculate, but there's still variability that can come from the availability and difficulty of mining the ore. There's also political issues to consider. If your country decides to shut down your nuclear plants prematurely, like Germany did, the huge upfront cost of building the plants can't be recouped by running them for an additional 10, 20, 30+ years until it becomes necessary to decommission them.

Regardless of how renewables are compared to nuclear, coal and gas are the elephants in the room when it comes to CO2 produced per kWh.